Zobrazeno 1 - 10
of 30
pro vyhledávání: '"Julian Salazar"'
Autor:
Valeria Franco Salvi, Julian Salazar, Jordi López Lillo, Agustina Vázquez Fiorani, Juan Montegú
Publikováno v:
Estudios Atacameños, Vol 69 (2023)
Los estudios realizados hasta el momento nos permiten concebir al valle de La Ciénega como un palimpsesto de prácticas humanas desde hace por lo menos 2500 años. La principal ocupación la registramos durante el primer milenio de la Era, aunque so
Externí odkaz:
https://doaj.org/article/fbf0bc1f269f4f238e237009e30b511b
Publikováno v:
Transactions of the Association for Computational Linguistics. 10:1249-1265
Large pretrained language models (PLMs) are often domain- or task-adapted via finetuning or prompting. Finetuning requires modifying all of the parameters and having enough data to avoid overfitting while prompting requires no training and few exampl
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::a2b27eb7c12c1d259264dd4e675653bf
https://doi.org/10.1162/tacl_a_00517
https://doi.org/10.1162/tacl_a_00517
Autor:
Hugo Díez, German Ramirez-Fo, Luisa Perez, Julian Salazar, Andrea Baracaldo, Carolina Ríos, Jorge Zambrano, Diana Castano, IvonneTorres, Diego Hernandez, Karen Sarmiento
Publikováno v:
Journal of Pharmacology and Toxicology. 15:8-15
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::493866c3e5be7be2fe789c24b1184f53
https://doi.org/10.3923/jpt.2020.8.15
https://doi.org/10.3923/jpt.2020.8.15
Publikováno v:
Odyssey
We introduce BERTphone, a Transformer encoder trained on large speech corpora that outputs phonetically-aware contextual representation vectors that can be used for both speaker and language recognition. This is accomplished by training on two object
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::48f32284d6bd716663a737ed3ddd7b77
https://doi.org/10.21437/odyssey.2020-2
https://doi.org/10.21437/odyssey.2020-2
We describe an unsupervised method to create pseudo-parallel corpora for machine translation (MT) from unaligned text. We use multilingual BERT to create source and target sentence embeddings for nearest-neighbor search and adapt the model via self-t
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::48fa85d532ed05079d727b6de4ab2c9d
Publikováno v:
EMNLP (1)
Multilingual contextual embeddings have demonstrated state-of-the-art performance in zero-shot cross-lingual transfer learning, where multilingual BERT is fine-tuned on one source language and evaluated on a different target language. However, publis
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::1190d9e9716622ae1ce72ba7bbe184af
https://doi.org/10.18653/v1/2020.emnlp-main.40
https://doi.org/10.18653/v1/2020.emnlp-main.40
Esta obra tiene como propósito dar a conocer las prácticas mortuorias, materialidades y tradiciones culturales de las poblaciones indígenas del Noroeste Argentino. Su desarrollo se sustenta en investigaciones arqueológicas y en las excavaciones i
Autor:
Karen Sarmiento, Julian Salazar, Patricia Hidalgo-Martínez, William Quevedo-Buitrago, Laura Ruiz, Carolina Ríos, Ariadna Rodríguez, Ivonne Torres, Hugo Díez
Publikováno v:
Universitas Médica. 61
En Colombia se presentan anualmente alrededor de 5000 casos de mordedura de serpiente y su tratamiento se basa en la neutralización con inmunoglobulinas completas purificadas, sin embargo, globalmente se utilizan antivenenos faboterápicos. Objetivo
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::7335b348c8c38353d19db7c2593f562d
https://doi.org/10.11144/javeriana.umed61-1.anti
https://doi.org/10.11144/javeriana.umed61-1.anti
Publikováno v:
ACL
Pretrained masked language models (MLMs) require finetuning for most NLP tasks. Instead, we evaluate MLMs out of the box via their pseudo-log-likelihood scores (PLLs), which are computed by masking tokens one by one. We show that PLLs outperform scor
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::d615161b81b34b1452fdd7c324692f7e
http://arxiv.org/abs/1910.14659
http://arxiv.org/abs/1910.14659
Publikováno v:
Web of Science
ICASSP
ICASSP
We propose a novel approach to semi-supervised automatic speech recognition (ASR). We first exploit a large amount of unlabeled audio data via representation learning, where we reconstruct a temporal slice of filterbank features from past and future
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::b30c06f26944045d7fe376e62418e145